A number of chemicals of interest have been tested for mutagenicity using Salmonella typhimurium in a number of different protocols. The rat liver carcinogen methapyriline was previously reported to be nonmutagenic in Salmonella. However, in this study, it produced a low-level, reproducible, dose-related mutagenic response in strain TA1535. The same response was also seen in two other laboratories. The mutagenic response seen was similar to the response seen with phenobarbital, another rodent liver carcinogen generally considered nongenotoxic. Phenobarbital is being studied, along with a series of chemicals having similar structures to portions of the phenobarbital molecule, in an attempt to elucidate its mutagenic mechanism. HC blue 1 and 2 are both direct-acting mutagens but only HC blue 1 is a rodent carcinogen. We obtained purified samples of HC blue 1 which have been reported to be nonmutagenic, but retain their carcinogenicity. A "cleaned-up" sample of the mutagenic HC blue 1 lost approximately 90% of its mutagenicity, but retained its carcinogenicity. Another carcinogenic sample that had been synthesized by a different process was not mutagenic. The human carcinogen treosulphan had not been previously reported as tested in Salmonella. In order to fill this data gap, it was tested in strains TA100 and TA1535, with and without metabolic activation. It was mutagenic in both strains under both activation conditions. Preliminary mutagenicity studies have been done on a crude preparation of red-tide (Ptychodiscus brevis) toxin. It was highly toxic to the Salmonella but not mutagenic. Additional studies have to be done to confine the nonmutagenicity and to extend the study to other Salmonella strains and metabolic activation systems.